diff --git a/pix2pixHD_attack/models/pix2pixHD_model.py b/pix2pixHD_attack/models/pix2pixHD_model.py
index dd99e48..daa2d62 100755
--- a/pix2pixHD_attack/models/pix2pixHD_model.py
+++ b/pix2pixHD_attack/models/pix2pixHD_model.py
@@ -241,7 +241,7 @@ class Pix2PixHDModel(BaseModel):
         else:
             input_concat = input_label  
 
-        input_adv = torch.clamp(input_concat + perturb * 1000, min=-1, max=1)   
+        input_adv = torch.clamp(input_concat + perturb, min=-1, max=1)   
            
         with torch.no_grad():
             fake_image = self.netG.forward(input_adv)
diff --git a/pix2pixHD_attack/test.py b/pix2pixHD_attack/test.py
index 6c5296c..198a538 100755
--- a/pix2pixHD_attack/test.py
+++ b/pix2pixHD_attack/test.py
@@ -52,14 +52,15 @@ for i, data in enumerate(dataset):
         exit(0)
     minibatch = 1 
 
-    if i == 0:
-        adv_image, perturb = model.attack(data['label'], data['inst'], data['image'])
+    # if i == 0:
+    #     adv_image, perturb = model.attack(data['label'], data['inst'], data['image'])
     if opt.engine:
         generated = run_trt_engine(opt.engine, minibatch, [data['label'], data['inst']])
     elif opt.onnx:
         generated = run_onnx(opt.onnx, opt.data_type, minibatch, [data['label'], data['inst']])
     else:        
         # generated = model.inference(data['label'], data['inst'], data['image'])
+        adv_image, perturb = model.attack(data['label'], data['inst'], data['image'])
         generated, adv_img = model.inference_attack(data['label'], data['inst'], data['image'], perturb)
         
     visuals = OrderedDict([('input_label', util.tensor2label(adv_img.data[0], opt.label_nc)),
diff --git a/pix2pixHD_attack/util/attacks.py b/pix2pixHD_attack/util/attacks.py
index 54db657..b6296ff 100644
--- a/pix2pixHD_attack/util/attacks.py
+++ b/pix2pixHD_attack/util/attacks.py
@@ -7,7 +7,7 @@ import torch
 import torch.nn as nn
 
 class LinfPGDAttack(object):
-    def __init__(self, model=None, epsilon=0.2, k=1, a=0.01):
+    def __init__(self, model=None, epsilon=0.05, k=1, a=0.05):
         self.model = model
         self.epsilon = epsilon
         self.k = k